Ferritin is a highly conserved 24 subunit protein that found in all animals, bacteria, and plants. The major physiological function of ferritin is to control the rate and location of polynuclear Fe(III)2O3 formation (see, e.g., Theil, E. C. “The ferritin family of iron storage proteins,” Adv. Enzymol. Relat. Areas Mol. Biol. 63:421-449 (1990), and Harrison, P. M., Lilley, T. H. “Ferritin in Iron Carriers and Iron Proteins,” Loehr T. M., ed. Weinheim: VCH, 1990:353-452; these and all references cited in the present application are incorporated herein by reference). This control is achieved through biomineralization which is performed by transporting hydrated iron ions and protons to and from a mineralized core. Through this mechanism, ferritin accumulates iron at concentrations orders of magnitude greater than the solubility of free iron under physiological conditions. The rate of biomineralization is directly related to the ratio of ferritin H and L subunits (the so-called heavy and light chains) within each capsid and exhibits the general trend of increasing the rate of iron storage with increasing H chain content. These differences in capsid composition are tissue dependent and affect the mechanism of iron oxidation, core formation and iron turnover. For example, ferritin comprised of predominantly L chain is found in the serum, while ferritin from the heart has a high ferritin H content. The ferritin mineralized iron core acts to provide bioavailable iron to a variety of redox enzymes and also serves a detoxification role.
Each ferritin protein is in the form of a 24 subunit capsid having 432 symmetry, a diameter of 125 Å, a shell thickness of approximately 25 Å and a hollow inner core of approximately 80 Å in diameter (FIG. 1). The monomeric ferritin typically has at least two isoforms denoted the L and H chains which differ in amino acid sequence. Although multiple forms of H and L subunit lengths have been identified in many vertebrates including humans, these two forms are generally both found in the ferritins that have been identified. Each ferritin subunit is approximately a 17 kilodalton protein having the topology of a helix bundle which includes a four-antiparallel helix motif, with a fifth shorter helix (the C-terminal helix) lying roughly perpendicular to the long axis of the 4 helix bundle. The helices are according to convention labeled ‘A, B, C, D & E’ from the N-terminus respectively. The N-terminal sequence lies adjacent to the capsid three-fold axis and clearly extends to the surface, while the E helices pack together at the four-fold axis with the C-terminus extending into the capsid core. The consequence of this packing creates two pores on the capsid surface. The pore at the four-fold is approximately 4 to 5 Å across and predominantly hydrophobic, while the three-fold pore, being slightly larger at 6.0 Å diameter is predominantly hydrophilic. It is expected that one or both of these pores represent the point by which the hydrated iron diffuses into and out of the capsid.
Previous work on ferritins, such as disclosed in U.S. Pat. Nos. 5,248,589; 5,358,722; and 5,304,382, all incorporated herein by reference, has focused on the physical aspects of the protein shell and the core such that materials other than ferrihydrate may be located inside the shell. It has also been shown (S P Martsev, A P Vlasov, P Arosio, Protein Engineering vol. 11, 377-381 (1998)) that recombinant human L and H ferritin when explored by differential scanning calorimetry will dissociate into subunit monomers at pH 2.0 to 2.8.
Other recent works have involved the use of “virus-like” particles as a modular system for vaccines wherein antibody responses were induced in the absence of adjuvants resulting in protection from viral infection and allergic reactions (Lechner et al., Intervirology 2002; 45(4-6); 212-7), but this system did not involved a ferritin-based development of proteins. In Marchenko et al., J. Mol. Microbiol Biotechnol 2003; 5(2):97-104, virus-like particles (VLPs) were constructed from a protein known as P1-380 which forms VLPs. In this case, fusion at the C and/or N-termini of the P1-380 protein did not interfere with the VLP self-assembly, and bi-functional fusion particles were made which demonstrated that they are more potent at generating and immune response. Still further, Douglas et al. have performed some work wherein a protein for the nucleation of iron was linked with the cowpea mosaic virus (CCMV), See Adv. Mater., 14 (6):415-418 (2002). Still other references refer to a “chimeric” protein using a virus-like particle which contains a nonstructural papillomavirus protein fused to the virus L2, a minor capsid protein. See Greenstone et al., PNAS USA, 95(4): 1800-5 (1998). However, in all of these cases, these fusion proteins did not involve ferritin.
Accordingly, none of the prior references have focused on utilizing ferritin or the placement of the N and C-termini at the outer and inner surface of the capsid respectively (e.g., as shown in FIGS. 2A & B, and described further below) for any purpose, and moreover, no one has previously has utilized this structure for the purpose of linking suitable proteins or peptides via fusion to ferritin in order to enhance the properties of the proteins or peptides while creating a fusion protein capable of self-assembly.